Re: checked in sieve code - correction to build instructions

[Sam Roberts]

Quoting Alain Magloire <address@hidden>, who wrote:
> > 
> > Unused, hardly, all caveats apply.
> > 
> > To build, add sieve/Makefile to the AC_OUTPUT statement
> > at about line 165 of configure.in.

And add sieve to the SUBDIRS line in the top-level Makefile.am.

> > -- off topic --
> > 
> > I suspect my mbox-explode.c example app of currupting and
> > truncating a large maibox of mine while running on it and
> > being SIGINTed, despite having opened it readonly. Also, it
> 
> In theory that is not possible, if you open the file readonly
> you simply can not write to it.  It is not mailutils/mailbox
> that is preventing this but the OS since the file was open (, O_RDONLY);
> and for mmap(, PROT_READ);
> So my guess is, the orginal mailbox was already corrupted
> by previous experiences and it was unnoticed.

I haven't been able to prove this, but I'm confused. It's was
a new mailbox, just written with the results of all the daily
tests. You are sure that the mailbox isn't opened writeable
briefly to update/add the Status and uid header fields?

The only other program I used was mutt to look at the mailbox.

I haven't been able to duplicate this, so I guess it's kind of
nebulous. I was hammering the disk pretty had, linux was basically
frozen, who knows what could have caused this.

> > I am trying to save, which may be me using the wrong apis,
> > or related to the mbox being corrupted, I'm trying to
> > figure out what happened.

The reason I was hammering the disk so hard was i was saving
the entiry msg as an attachment, instead of the part, oops. I'll
patch the mbox-explode.c soon. But first, spring cleaning.

Sam

-- 
Sam Roberts <address@hidden> (Vivez sans temps mort!)